Nondestructive assessment of eggshell thickness by VIS/NIR spectroscopy

Abstract. The objective of this research was to investigate the feasibility of transmission spectroscopy for the non-destructive assessment of eggshell thickness. A total of 70 intact white-shelled eggs of the same flock were measured. Pretreatment methods of standard normal variate, multiplicative scatter correction, the first derivative and the second derivative were applied for the original spectrum. A partial least squares model was built in order to predict eggshell thickness. The results illustrated partial least squares regression after multiplicative scatter correction pretreatment model for whole egg pH was superior to the other methods. Correlation coefficients were 0.86, 0.84 for calibration and prediction model, respectively. Standard errors were both 0.01 for calibration model and prediction model. This study illustrated visible near-infrared spectroscopy and partial least squares regression were useful for nondestructive assessment of eggshell thickness.